Field of the Invention
The present invention relates to color transformation processing of image data input from an image input device.
Description of the Related Art
Spectral sensitivity characteristics of an image input device as typified by a digital camera commonly do not match visual spectral characteristics (a color matching function) of a person. An RGB value of image data acquired by an image input device is a value that depends on characteristics of the image input device. Therefore, it is necessary to transform an RGB value that is a device-dependent color signal (hereinafter referred to as a device RGB value) into tristimulus values (XYZ values) that are device-independent color signals or a standard RGB value such as in an sRGB space or an AdobeRGB space.
A method for transforming a device RGB value into tristimulus values by a matrix operation is known, but by matrix operation processing a transformation having high precision high chroma color in particular is not achieved, and a color transformation precision decreases. Accordingly, the International Publication 2013/101639 discloses an invention that performs high precision color transformation processing in the entire region of visible light by performing a transformation from a device RGB value to an XYZ value by a lookup table (LUT) computation instead of a matrix operation.
In the invention of International Publication 2013/101639, to generate an LUT used in the color transformation, grid partitioning is performed for the entirety of a visible light range, and spectral data that corresponds to each grid point is generated. Subsequently, a device RGB value and an XYZ value corresponding to the spectral data of each grid point is calculated, and an LUT indicating a correspondence between device RGB values and XYZ values is formed. This LUT has a two-dimensional input/output format, and the invention of the International Publication 2013/101639 transforms three-dimensional color information (RGB values) into two-dimensional chromaticity information p=R/(R+G+B), q=G/(R+G+B), and inputs the chromaticity information pq into the LUT to obtain two-dimensional xy chromaticity data. The xy chromaticity data is then used to calculate tristimulus values (XYZ values).
To transform xy chromaticity data to XYZ values, conventionally there is a necessity for a value calculated by (X+Y+Z) as a transformation coefficient. Because the invention of the International Publication 2013/101639 uses a value calculated by (R+G+B) as an alternative for a transformation coefficient to calculate the XYZ values, it cannot be said that calculation precision of the XYZ values is high.